News

  • [Jan 2024] : My paper on mitigating system-level failures in visual controllers got accepted into ICRA 2024.
  • [Jan 2024] : Selected to attend Research Week with Google.
  • [Nov 2023] : New paper on drone docking on offshore platforms using deep-RL submitted to IEEE Transactions on Artificial Intelligence.
  • [Sept 2023] : New paper on mitigating system-level failures in visual controllers is out on arxiv.
  • [May 2023] : Started Research Internship at SIA Lab, University of Southern California.
  • [Jan 2023] : Started Research Internship at RNCSL, Carleton University.
  • [Dec 2022] : Appointed as Tech Lead of RoboReG (Robotics Research Group).
  • [May 2022] : Started Research Internship at RBCCPS, IISc.
  • [Mar 2022] : Started Research Internship at AISL, CWNU
  • [Dec 2021] : First Runner Up in AIITRA Robotics Challenge
  • [Nov 2020] : Started B.Tech in Electronics Engineering at IIT BHU

Publications

Detecting and Mitigating System-Level Anomalies of Vision-Based Controllers

Aryaman Gupta, Kaustav Chakraborty, Somil Bansal
IEEE International Conference on Robotics and Automation (ICRA) 2024

Deep Reinforcement Learning for Sim2Real Policy Transfer in VTOL-UAVs Offshore Docking Operations

Aryaman Gupta, Ali Mohamed Ali, Hashim A.Hashim

Experience

University of Southern California

Research Intern | Prof. Somil Bansal | May 2023 - Present
Working on mitigating system-level failures of vision-based controllers in aircraft taxiing and ground robot navigation tasks.

Carleton University

Research Intern | Dr. Hashim Mohamed | Jan 2023 - Dec 2023
Working on developing a deep-RL based docking mechanism for VTOL UAVs on offshore platforms.

Indian Institute of Science

Research Intern | Dr. Bharadwaj Amrutur | May 2022 - July 2022
Worked on multi-agent exploration using RRT-based methods and dynamic obstacle avoidance using 3D object detection.

Changwon National University

Research Intern | Dr. Oh-Seol Kwon | Mar 2022 - July 2022
Worked on combining Faster R-CNN and Edge Enhanced SRGAN architectures for enhanced object detection in low resolution aerial images.

Projects

Multi-Agent Warehouse Coverage & Cleaning

Efficient coverage and cleaning of unknown terrains using multiple ground robots.

Mult-Agent Exploration & Dynamic Obstacle Avoidance

RRT-Exploration and visual obstacle detection and avoidance package.

UAV Control and Swarming

Cascaded PID-based control and swarm motion of drones.

Multi-Purpose Household Robot

Fabricated a ground robot to perform household chores and child care.

Small World IoT Networks

Enhanced network performance by introducing small-world phenomenon using actor-critic reinforcement learning.

UAV Guided UGV Navigation

UAV aided mapping and localization for a UGV to autonomously traverse in mountainous terrains.